Aluminium is produced conventionally by the Hall-Heroult process, by the electrolysis of alumina dissolved in cryolite-based molten electrolytes at temperatures up to around 950.degree. C. A Hall-Heroult reduction cell typically has a steel shell provided with an insulating lining of refractory material, which in turn has a lining of carbon which contacts the molten constituents. Conductor bars connected to the negative pole of a direct current source are embedded in the carbon cathode substrate forming the cell bottom floor. The cathode substrate is usually an anthracite based carbon lining made of prebaked cathode blocks, joined with a ramming mixture of anthracite, coke, and coal tar.
In Hall-Heroult cells, a molten aluminium pool acts as the cathode. The carbon lining or cathode material has a useful life of three to eight years, or even less under adverse conditions. The deterioration of the cathode bottom is due to erosion and penetration of electrolyte and liquid aluminium as well as intercalation of sodium, which causes swelling and deformation of the cathode carbon blocks and ramming mix. In addition, the penetration of sodium species and other ingredients of cryolite or air leads to the formation of toxic compounds including cyanides.
The anodes are usually prebaked carbon blocks that are consumed by the electrochemical reaction, corroded by contact with the electrolyte and eroded by the evolved oxidizing gases. Attempts to replace these consumable carbon anodes by non-consumable oxygen evolving anodes have encountered difficulties linked to the inherent porosity of the ceramic and metal ceramic materials proposed for this purpose, and because of the imperfect impermeability of protective coatings, e.g. based on cerium oxyfluoride, whereby unwanted species penetrate into the bulk of the anode material and cause damage.
The problems associated with penetration of sodium into the carbon cathode have been extensively studied and discussed in the literature.
Several papers in Light Metals 1992 published by the The Minerals, Metals and Materials Society discuss these problems. A paper "Sodium, Its Influence on Cathode Life in Theory and Practice" by Mirtag et al, page 789, emphasises the advantages of using graphitic carbon over anthracite. Reasons for the superiority of graphitic carbon were also set out in a paper "Change of the Physical Properties and The Structure in Carbon Materials under Electrolysis Test" by Ozaki et al, page 759. Another paper "Sodium and Bath Penetration into TiB.sub.2 Carbon Cathodes During Laboratory Aluminium Electrolysis" by Xue et al, page 773, presented results showing that the velocity of sodium penetration increased with increasing TiB.sub.2 content. Another paper "Laboratory Testing of the Expansion Under Pressure due to Sodium Intercalation in Carbon Cathode Materials for Aluminium Smelters" by Peyneau et al, page 801, also discusses these problems and describes methods of measuring the carbon expansion due to intercalation.
There have been several attempts to avoid or reduce the problems associated with the intercalation of sodium in carbon cathodes in aluminum production.
Some proposals have been made to dispense with carbon and instead use a cell bottom made entirely of alumina or a similar refractory material, with a cathode current supply arrangement employing composite current feeders using metals and refractory hard materials. See for example, EP-B-0 145 412, EP-A-0 215 555, EP-B-0 145 411, and EP-A-0 215 590. So far, commercialization of these promising designs has been hindered due to the high cost of the refractory hard materials and difficulties in producing large pieces of such materials.
Other proposals have been made to re-design the cell bottom making use of alumina or similar refractory material in such a way as to minimize the amount of carbon used for the cathode--see U.S. Pat. Nos. 5,071,533 and 5,135,621. Using these designs will reduce the problems associated with carbon, but the carbon is still subject to attack by sodium during cell start up.
There have been numerous proposals to improve the carbon materials by combining them with TiB.sub.2 or other refractory hard materials, see e.g. U.S. Pat. No. 4,466,996. But, as pointed out in the above-mentioned paper of Xue et al, with such composite materials, the penetration increases with increasing TiB.sub.2 content.
WO/93/20027 proposes applying a protective coating of refractory material to a carbon cathode by applying a micropyretic reaction layer from a slurry containing particulate reactants in a colloidal carrier, and initiating a micropyretic reaction. To assist rapid wetting of the cathode by molten aluminium, it was proposed to expose the coated cathode to a flux of molten aluminium containing a fluoride, a chloride or a borate of lithium and/or sodium. This improves the wetting of the cathode by molten aluminium, but does not address the problem of sodium attack on the carbon, which is liable to be increased due to the presence of TiB.sub.2.
No adequate solution has yet been proposed to substantially reduce or eliminate the problems associated with sodium penetration in carbon cathodes, namely swelling especially during cell start-up, displacement of the carbon blocks leading to inefficiency, reduced lifetime of the cell, the production of large quantities of toxic products that must be disposed of when the cell has to be overhauled, and the impossibility to use low density carbon. Corresponding problems of penetration of unwanted species into anodes also remain unsolved.